Struct linfa_clustering::AppxDbscanLabeler

pub struct AppxDbscanLabeler { /* fields omitted */ }

Struct that labels a set of points according to the Approximated DBSCAN algorithm

Implementations

impl AppxDbscanLabeler

pub fn new<F: Float>(
    observations: &ArrayView2<'_, F>,
    params: &AppxDbscanValidParams<F>
) -> AppxDbscanLabeler

Runs the Approximated DBSCAN algorithm on the provided observations using the params specified in input. The Labeler struct returned contains the label of every point in observations.

Parameters:

• observations: the points that you want to cluster according to the approximated DBSCAN rule;
• params: the parameters for the approximated DBSCAN algorithm

Return

Struct of type Labeler which contains the label associated with each point in observations

pub fn labels(&self) -> &Array1<Option<usize>>

Gives the labels of every point provided in input to the constructor.

Example:

use ndarray::{array, Axis};
use linfa::prelude::*;
use linfa_clustering::{AppxDbscanLabeler, AppxDbscan};

// Let's define some observations and set the desired params
let observations = array![[0.,0.], [1., 0.], [0., 1.]];
let params = AppxDbscan::params(2).check().unwrap();
// Now we build the labels for each observation using the Labeler struct
let labeler = AppxDbscanLabeler::new(&observations.view(),&params);
// Here we can access the labels for each point `observations`
for (i, point) in observations.axis_iter(Axis(0)).enumerate() {
    let label_for_point = labeler.labels()[i];
}